Chase Walker

PhD researcher in Explainable AI (XAI) at the University of Florida (anticipated graduation August 2026).

My work centers on a critical question in modern AI: Can we have state-of-the-art high-performance models that are not only accurate, but can also be interpreted, provably robust, and safe for real-world use?

To answer this question, my research contributions include developing novel attribution methods, evaluation metrics, and XAI-driven adversarial detection systems to improve model transparency and reliability. My work has been published in leading AI and ML conferences, including AAAI, ICLR, IJCAI, and AISTATS, with topics spanning:

• Attribution methods for CNNS, ViTs, and LLMs
• Model agnostic and specific attribution evaluation metrics
• XAI-based robustness and confidence measures
• Neuro-symbolic AI

In addition to technical research, I mentor undergraduate teams, deliver invited talks, and build reproducible PyTorch pipelines for large-scale experimentation. My work bridges academic research and practical deployment, emphasizing clarity, scientific rigor, and human-aligned AI systems.

Outside of work, I enjoy playing guitar, exploring new video games, reading, and traveling. I’m especially drawn to learning about other cultures through their food, art, and day-to-day life experiences.